The probability-problem study illustrated the effects of semantic similarity on the analogical mapping process. In a number of studies, Ross (1987, 1989) had subject solve various types of probability problems using sample problems to guide their solution. The studies included five experimental conditions based on the similarity of story lines and corresponding objects, categorized by the notation {+/0}/{+/Ð/0}. The first symbol in this notation represents the similarity of story lines, where + indicates similar story lines and 0 represents dissimilar story lines. The second symbol represents the similarity of corresponding objects, where + indicates that similar objects correspond, Ð indicates that similar objects do not correspond, and 0 indicates that similarities are neutral with respect to correspondences. When examining five conditions (+/+, +/Ð, 0/+, 0/Ð, and 0/0), Ross showed that semantic similarity can positively influence analogical mapping when similar objects correspond and can negatively influence mapping when similar objects do not correspond.
The probability-problem model simulates performance for each of the five conditions. In each simulation, the model maps the quantity cars in the sample problem to an analogous quantity in the test problem. The results show the proportion correct for each condition.